Identifying disease genes is an important step towards effective treatment of human disease. Efforts to identify human disease genes are both cursed and blessed by the existence of population structure, which can lead to false disease associations in some disease mapping approaches but actually drives the existence of others. This research proposal aims to alleviate these curses and harness these blessings. The first aim is to improve upon existing methods to correct whole-genome haplotype association studies for population stratification, i.e. systematic ancestry differences between cases and controls, which is a significant source of false disease associations. The second aim is to apply whole-genome ancestry association mapping to Latino populations; the focus will be on type II diabetes, which is particularly prevalent in Native Americans and may thus induce an elevated proportion of Native American ancestry in Latinos with type II diabetes near a disease gene. The third aim is to develop a method for whole-genome natural selection mapping which harnesses the differentiation between sampled populations to identify genes under recent natural selection, which are strong candidates for relevance to disease. [unreadable] [unreadable] [unreadable]